Response surface-guided multi-objective optimization of CR/NR blend mechanical and rheological properties using artificial bee colony algorithm

In this study, Response Surface Methodology and the Artificial Bee Colony (ABC) algorithm were employed to optimize the vulcanization characteristics and mechanical properties of Chloroprene Rubber (CR) and Natural Rubber (NR) blends. CR is a commonly used but expensive material in the automotive industry. Therefore, to achieve an optimal balance between cost and performance, this study explored the feasibility of using CR/NR blends in industrial applications. The effects of accelerators, retarders, vulcanization agents, and the proportion of NR in the final blend on the response variables were systematically investigated. In the first stage of the study, experiments were conducted based on a Central Composite Design. Following Analysis of Variance-ANOVA, regression models were developed for the response variables: hardness, tensile strength, and Tan δ. In the second stage, the conflicting regression models were treated as objective functions within a multi-objective optimization frameworkutilizing the ABC algorithm. These objectives were scalarized using the weighted sum, conic, and Tchebycheff scalarization methods. Among the solutions obtained using the Conic Scalarization Method (CSM), the sensitivity analysis revealed that the trade-off behavior can be effectively tuned by adjusting objective weights. When Tan δ minimization was prioritized (e.g., w₃ = 0.8), the algorithm achieved a Tan δ value as low as 0.076, with a corresponding hardness of 49.53 ShA and tensile strength of 16.23 MPa. Conversely, when mechanical performance was emphasized (e.g., w₁ = 0.8), a maximum hardness of 57.15 ShA and tensile strength of 19.38 MPa were obtained, with Tan δ of 0.198. When tensile strength was dominant (e.g., w₂ = 0.8), the tensile strength reached 24.86 MPa, accompanied by 53.60 ShA hardness and 0.238 Tan δ. These results demonstrate that the CSM-based MOABC approach offers robust and tunable optimization performance, providing engineers with the flexibility to adjust formulations according to specific design priorities in CR/NR-based automotive rubber components.